Advancing Event Causality Identification via Heuristic Semantic Dependency Inquiry Network

TL;DR

This paper introduces SemDI, a novel neural network model that improves event causality identification by capturing semantic dependencies and using a Cloze Analyzer for better context understanding, outperforming existing methods.

Contribution

The paper presents SemDI, a new approach that leverages semantic dependency inquiry and a Cloze Analyzer to enhance causal relation detection in texts, addressing limitations of previous reliance on explicit features and external knowledge.

Findings

SemDI outperforms state-of-the-art methods on three benchmarks.

Semantic dependency capture improves causal relation identification.

The Cloze Analyzer effectively models context for causality inference.

Abstract

Event Causality Identification (ECI) focuses on extracting causal relations between events in texts. Existing methods for ECI primarily rely on causal features and external knowledge. However, these approaches fall short in two dimensions: (1) causal features between events in a text often lack explicit clues, and (2) external knowledge may introduce bias, while specific problems require tailored analyses. To address these issues, we propose SemDI - a simple and effective Semantic Dependency Inquiry Network for ECI. SemDI captures semantic dependencies within the context using a unified encoder. Then, it utilizes a Cloze Analyzer to generate a fill-in token based on comprehensive context understanding. Finally, this fill-in token is used to inquire about the causal relation between two events. Extensive experiments demonstrate the effectiveness of SemDI, surpassing state-of-the-art methods on three widely used benchmarks. Code is available at https://github.com/hrlics/SemDI.